Skip directly to searchSkip directly to the site navigationSkip directly to the page's main content

Evidence-Based Public Health Practice

"Evidence-based" is a term in popular use in public health and medicine today, and with good reason - evidence-based interventions should provide the best outcomes for patients and populations.

The following definition for evidence-based public health practice that is currently being used by some in the New Mexico Department of Health.

Evidence-based public health practice is the careful, intentional and sensible use of current best scientific evidence in making decisions about the choice and application of public health interventions.

1 Examples of Evidence-Based Practice
2 Strength of Evidence
3 Where do I find Evidence-Based Practices?
4 Research Methodologies


Back to Top

Examples of Evidence-Based Practice

Here are two examples that may help in understanding what evidence-based means:


The "evidence" is the reduction in DWI, and in this case, the evidence suggests that the intervention was effective. Although another aspect to consider is whether there were there other factors that could account for the reduction in DWI (e.g., road checks, public education campaign).


The "evidence" in this example is the decrease in bullying following the intervention. Other factors to consider: where was the school (in the USA or Canada, in a wealthy or low income part of town?) Since the evaluation was conducted with students in grades 3-6, can one assume the intervention will work with the older, middle-school, students?

Back to Top

Strength of Evidence

As you might imagine, not all evidence is created equal. Here is a short list of what to look for in evaluating evidence for best practices in public health. The most desirable interventions are those that maximize each of the following criteria.

Characteristics of the Evidence:GoodPoor
Volume of Evidence: Has the study been repeated, or replicated by different researchers and in different populations?REPLICATED IN MULTIPLE STUDIESWHAT STUDIES?
Quality of Evidence: Does the evidence derive from research studies that were designed to establish a cause-and-effect relationship between the intervention and the outcome? Usually (but not always), studies published in peer-reviewed journals will utilize a credible study design.CAUSAL, RANDOMIZED, CONTROLLED STUDY DESIGNANECDOTES, CASE STUDIES, SPECULATION
Generalizability: Can you generalize from the study population to your local population? For instance, a study of Australian college students won't generalize to New Mexico 3rd graders.STUDY POPULATION SIMILAR TO LOCAL POPULATIONSTUDY POPULATION DIFFERENT FROM LOCAL POPULATION

Characteristics of the Intervention:GoodPoor
Time from intervention to outcome: How long after you've implemented your intervention should you expect to see results?SHORT INTERVAL (LESS THAN 4 YEARS)INTERVAL LONG OR UNKNOWN
Sustained Impact: How long after you've implemented your intervention should you expect to see results?LONG-LASTING EFFECT EFFECT FADES QUICKLY

Additional criteria may also be relevant, but the criteria above will help as you evaulate the evidence in favor or against potential interventions.

Back to Top

Where do I find Evidence-Based Practices?

These Websites, below, also contain useful guidance on specific evidence-based interventions.

Back to Top

Research Methodologies for Cause-and-effect Relationships

These three criteria must be met for a study to demonstrate a cause-and-effect relationship:
  1. Observed Statistical Association. There must be some statistical evidence of association between the cause and the effect.
  2. Time Precedence. The cause must occur first, followed by the effect.
  3. Rule out Alternative Explanations for the Association.

The last criterion is the most difficult to satisfy. A "true experiment" is a study design that is intended to rule out alternative explanations.

By definition, a "True Experiment" has the following characteristics.
  • The researcher must have a study group and a control group.
  • He or she must randomly assign participants to the study and control groups.
  • And he or she must manipulate an "independent variable" in the study group.

  • Without a true experiment, it is very difficult to establish a credible causal relationship.



    --Back to Glossary/Index Page--

    Please feel free to contact us with questions or suggestions.